Clinical and testing laboratories regularly analyse the same control material as many other labs and compare results. The standard deviation index (SDI) condenses that comparison into one number: how many peer-group standard deviations your lab’s mean lies away from the consensus mean. It is essentially a z-score for your laboratory.
An SDI of 0 means your results match the group exactly. A positive SDI means you read high, a negative one that you read low. Values within ±1 indicate good agreement, up to about ±1.25 is generally acceptable, values between 1.25 and 2 call for a closer look at calibration and trends, and an SDI of ±2 or more is usually considered unacceptable and needs corrective action. Watching SDI over several months reveals a developing bias long before a single result fails.
SDI looks at accuracy (bias). Precision is judged by the coefficient of variation ratio (CVR, sometimes CVI): your lab’s CV divided by the group’s CV. A CVR above 1 means your imprecision is larger than that of your peers.
How to use the Standard Deviation Index Calculator
- Laboratory mean
- Consensus group mean
- Consensus group standard deviation
- Also compare precision?
- Laboratory standard deviation
Results update as you type, and everything is calculated in your browser.
How it works
Enter your laboratory’s mean for the control, the consensus (peer-group) mean and the consensus standard deviation. Optionally add your lab’s standard deviation to get the CV ratio. SDI = (lab mean − group mean) ÷ group SD.
Worked example
A laboratory’s monthly mean for a glucose control is 5.32 mmol/L, while the peer group reports 5.10 mmol/L with SD 0.16 mmol/L. SDI = (5.32 − 5.10) ÷ 0.16 = 1.375 – marginal, the lab reads high by 0.22 mmol/L (4.3%). The lab’s own SD is 0.14, so its CV is 2.63% against the group’s 3.14%, and CVR = 0.84: precision is fine, the issue is calibration bias.
SDI limits
| |SDI| | Interpretation | Typical action |
|---|---|---|
| 0 – 1.0 | Good | None |
| 1.0 – 1.25 | Acceptable | Monitor |
| 1.25 – 2.0 | Marginal | Review calibration, reagents, trends |
| ≥ 2.0 | Unacceptable | Investigate and correct |
Inputs explained
| Field | What to enter |
|---|---|
| Laboratory mean | A number |
| Consensus group mean | A number |
| Consensus group standard deviation | A number |
| Also compare precision? | Choose one: No, Yes – enter the laboratory SD |
| Laboratory standard deviation | A number |
What each result means
- Standard deviation index (SDI) = ([Laboratory mean] – [Consensus group mean])÷[Consensus group standard deviation]
- Bias (lab − consensus) = [Laboratory mean] – [Consensus group mean]
- Bias as % of consensus mean = ([Laboratory mean] – [Consensus group mean])÷[Consensus group mean]×100
- Laboratory CV = [Laboratory standard deviation]÷[Laboratory mean]×100
- Consensus CV = [Consensus group standard deviation]÷[Consensus group mean]×100
- Coefficient of variation ratio (CVR) = [Laboratory CV]÷[Consensus CV]
Formula
SDI = (laboratory mean − consensus mean) ÷ consensus SD · CVR = lab CV ÷ consensus CV · bias % = (lab mean − consensus mean) ÷ consensus mean × 100
Worked example
Inputs — Laboratory mean: 5.32; Consensus group mean: 5.1; Consensus group standard deviation: 0.16; Also compare precision?: Yes – enter the laboratory SD; Laboratory standard deviation: 0.14.
Results — Standard deviation index (SDI) = 1.375; Performance = Marginal – check calibration and trends (1.25 – 2); Direction = Laboratory reads high; Bias (lab − consensus) = 0.22; Bias as % of consensus mean = 4.314 %; Laboratory CV = 2.632 %; Consensus CV = 3.137 %; Coefficient of variation ratio (CVR) = 0.839.
Key terms
- Consensus (peer-group) mean
- The average result of all laboratories using the same method and instrument.
- SDI
- The distance between the lab mean and the consensus mean in consensus standard deviations.
- CVR
- The laboratory's coefficient of variation divided by the peer group's.
Frequently asked questions
What is a good SDI?
0 is ideal. |SDI| ≤ 1 shows good agreement with peers, up to 1.25 is acceptable, 1.25–2 warrants investigation and 2 or more is unacceptable.
How do I calculate SDI?
Subtract the consensus mean from your lab mean and divide by the consensus SD. A lab mean of 9 against a group mean of 8 with SD 2 gives SDI = 0.5.
What does a negative SDI mean?
Your laboratory reads lower than the peer group. The size of the problem is the absolute value.
What is the difference between SDI and CVR?
SDI measures bias (accuracy) relative to peers; CVR compares your imprecision with theirs. CVR below 1 means you are more precise than the group.
What CV ratio is acceptable?
A CVR up to about 1.0 is good, 1.0–1.5 acceptable, 1.5–2.0 needs investigation and above 2.0 requires action.
Is SDI the same as a z-score?
Yes in form – it is a z-score of your lab's mean within the distribution of peer means. Proficiency-testing schemes often report z-scores with the same ±2 and ±3 limits.
How often should SDI be reviewed?
Monthly for each analyte and control level is common. Look for trends: three consecutive months moving the same way can signal a shift before limits are breached.